May I assume that Skritter usually tries to use the mainland style for traditional characters (not Taiwan or Hong Kong)?

Firstly, that's what I observe for the vast majority of traditional characters （even if there's no 草字头 ...）.

Secondly, any deviation from that rule would cause styles to vary across traditional characters and characters where no simplified version exists.

Still, quite a few characters still use the TW style (largely intuitive; I also checked on zdic.net but not on the website of the TW administration):

Some examples:

Example 1: 糨 - TW style.
http://www.zdic.net/zd/zi/ZdicE7ZdicB3ZdicA8.htm

Example 2: 扉 - TW style.
http://www.zdic.net/zd/zi/ZdicE6Zdic89Zdic89.htm

Example 3: 甄 vs. 湮
Same component, but 甄 (http://www.zdic.net/zd/zi/ZdicE7Zdic94Zdic84.htm) has the Taiwan style, 湮 (http://www.zdic.net/zd/zi/ZdicE6ZdicB9ZdicAE.htm) has the mainland style.

Example 4: 遞 vs. its radical 虒
遞 has mainland style, its radical 虒 has TW style, confused me quite a bit at first.

Example 5: 賴 and 懶 - TW style

=========

ADDENDUM:

To check styles, look up the character on zdic.net, then click on "字源字形". You'll see a 字形对比.

The 麻 component used to be written both in TW and in mainland style, depending on the character.

Now mainland style appears to be used consistently, at least I don't think I've been confronted with a TW-styled 麻 for the last couple of weeks.

deleted ( by self )

Our revealed policy up to this point has been something like, stick to the mainland style for components that are shared with simplified characters (most of them), but allow the Taiwan style for the rest because it angers Bohan whenever he sees one that could be Taiwan-style but isn't.

I guess we could switch over to mainland style entirely, but it's not a great solution in that mainland style just isn't as useful as Taiwan style for most learners of traditional characters--it's just what we're forced to use for compatibility with simplified characters.

Makes sense, Nick.

Please take a look the findings above nonetheless. Some of these are clear inconsistencies or even simplified characters with TW syle.

The more I study, the more revisions are spent on inconsistencies.

It is certainly nice to know the alternative traditional style of the 真 component. Being introduced to it through the traditional 癲 while all other characters with that component stay the same, is confusing, however.

The same goes for the alternative top radical of 窯 and the alternative traditional 麻 radical in 嬤.